A Milestone for Ecosystem Research: Artificial Intelligence in Wildlife Monitoring

Within the project, Prof. Dr. Manuel Steinbauer, Professor of Sports Ecology, is leading the research work at the University of Bayreuth. His team explores the increasing popularity of outdoor sports and its potential consequences on the environment, landscape, and wildlife. One key question is the extent to which wildlife is affected when natural habitats are increasingly used for sporting activities, such as mountain biking. “Digital technologies have created new opportunities for wildlife research and cross-regional wildlife management. This new project aims to fully leverage these possibilities. Through the automated evaluation of images using artificial intelligence and the resulting long observation time series, researchers can obtain detailed insights into how wild animals respond to human influences in their environment," says Steinbauer.

The Veldensteiner Forst, a forest in Bavaria, is home to a diverse range of animal species that are representative for the region. To monitor the forest's wildlife, a large-scale network of cameras has been in operation for several years, and it will soon be expanded to include all wildilfe species living in the forest. In addition to monitoring wildlife, visible changes in vegetation and weather will also be evaluated to provide a comprehensive understanding of the forest's ecosystem.

A systematic monitoring of wild animals using a network of cameras, a so-called "photo trap monitoring", generates very large amounts of image data. The data analysis, which aims to identify, among other things, the species and gender of the animals depicted, requires a high amount of effort. Within the new project, the evaluation of the image data is largely taken over by artificial intelligence. A team at LMU Munich will develop the necessary algorithms for this purpose. Data protection will also be taken into account: images of people will be automatically sorted out without losing the necessary information about human activities. This technological basis will allow new image-based insights into the interaction of humans and animals that is applicable worldwide. Results will promote proactive management of outdoor sports in landscapes that play an important role in nature conservation and the preservation of biodiversity.

Project partners:

Prof. Dr. Manuel Steinbauer is leading the project at the University of Bayreuth. His team explores the increasing popularity of outdoor sports and its potential consequences on the environment, landscape, and wildlife. Of particular interest is the extent to which wildlife is affected when natural habitats are used for outdoor sport.

Dr. Ludwig Bothmann works as a post-doc at the Chair of Statistical Learning and Data Science of the Institute of Statistics at LMU Munich. As part of the project, he and his team will train neural networks that have already been developed on new animal species and improve their accuracy in classification. In doing so, he can draw on his many years of experience in AI-based classification of image data as well as scientific evaluation of the accuracy and interpretability of machine learning algorithms.

In the project, Dr. Wibke Peters and Dr. Hendrik Edelhoff represent the section Wildlife Biology and Wildlife Management of the LWF in Freising. Their research deals with scientific assessment of wildlife populations, for example with regard to their condition, size and also use of space. They can build on extensive experience from several projects regarding the use of photo traps in wildlife monitoring. They now bring this expertise to the new project and want to further optimize the applicability of this methodology for practice and future research and investigate wildlife ecological relationships in the Veldenstein Forest.